Atherosclerosis xxx (2012) 1e7. Contents lists available at SciVerse ScienceDirect. Atherosclerosis

Transcription

1 Atherosclerosis xxx (2012) 1e7 Contents lists available at SciVerse ScienceDirect Atherosclerosis journal homepage: Histopathological correlates of the napkin-ring sign plaque in coronary CT angiography Harald Seifarth a,b, Christopher L. Schlett a, Masataka Nakano c, Fumiyuki Otsuka c, Mihály Károlyi a, Gary Liew a, Pál Maurovich-Horvat a, Hatem Alkadhi a, Renu Virmani c, Udo Hoffmann a, * a Cardiac MR PET CT Program, Massachusetts General Hospital and Harvard Medical School, 165 Cambridge Street, Boston, MA 02114, USA b Department of Clinical Radiology, University of Muenster, Germany c CV Path Institute Inc., Gaithersburg, Maryland, USA article info abstract Article history: Received 13 March 2012 Received in revised form 30 May 2012 Accepted 14 June 2012 Available online xxx Keywords: Coronary CT angiography Vulnerable plaque Atherosclerosis Angiogenesis Calcification Objective: The purpose of this study was to identify histologic characteristics of advanced coronary atherosclerotic plaques that are related with the detection of the napkin-ring sign (NRS) in coronary CT angiography (CCTA). Methods: CCTA was performed in 7 human donor hearts. Histological slicing and stainings were performed in 1 mm increments of each major coronary artery. Histology was co-registered with the CT-data and classified according to the modified AHA classification. Results: Advanced plaques (types IVeVI) were present in 139 (23%) of 611 cross sections. Of these 33 (24%) demonstrated an NRS in CCTA. NRS plaques were associated with greater non-core plaque area (median 10.2 vs. 6.4 mm 2, p < 0.01) and larger vessel area (median 17.1 vs mm 2, p < 0.01). The area of the necrotic/lipid core was larger in plaques with NRS (median 1.1 vs. 0.5 mm 2, p ¼ 0.05). Angiogenesis tended to be more frequent in plaques with NRS (48% vs. 30%) whereas micro-calcifications tended to be more frequent in plaques without NRS (27% vs. 46%) (p ¼ 0.06 and 0.07 respectively). In a multivariate analysis, necrotic/lipid core area (OR ¼ 1.9), non-core plaque area (OR ¼ 1.6), and total vessel area (OR ¼ 0.9) independently predicted the appearance of the NRS in coronary CT angiography. Conclusion: Delineation of NRS in CCTA is independently linked to the size of the necrotic/lipid core, the size of the non-core plaque and to the vessel area as measured in histology of advanced coronary atherosclerotic plaques. Ó 2012 Elsevier Ireland Ltd. All rights reserved. 1. Introduction Acute coronary syndromes frequently occur in patients with coronary atherosclerotic plaques exhibiting distinct morphological features such as increased plaque dimensions, a large lipid-rich core and a thin fibrous cap [1,2]. Initially described in histopathology from patients who died of acute myocardial infarctions, atherosclerotic plaques with similar features are also present in patients who have not experienced an acute cardiovascular event [3]. To date invasive techniques are available to detect plaque features associated with acute coronary syndromes [4,5]. Coronary CT angiography (CCTA) has emerged as a non-invasive tool to detect coronary artery stenosis and a growing body of * Corresponding author. Tel.: þ ; fax: þ address: uhoffmann@partners.org (U. Hoffmann). evidence suggests that it can also be used to discriminate fibrotic and lipid-rich components of coronary plaques based on differences in attenuation [6,7]. However because of a substantial overlap of the CT numbers of the various plaque types, a reliable differentiation of plaques is not possible with sufficient accuracy [8]. Recently, several small studies suggested that a certain morphological pattern in CT, a hyperdense ring-like structure surrounding a hypodense center, described as napkin-ring sign, (NRS) is associated with advanced coronary plaque as defined by the AHA classification [9e11]. Previous studies have shown that the napkin ring sign in CCTA offers a high specificity and positive predictive value to detect lipidrich plaques, yet the sensitivity of this sign to detect advanced plaque is rather low [9,12,13]. The purpose of this ex-vivo study was to determine histopathological features of advanced coronary atherosclerotic plaques that promote or interfere with the delineation of the napkin-ring sign in coronary CT angiography using co-registered data /$ e see front matter Ó 2012 Elsevier Ireland Ltd. All rights reserved.

2 2 H. Seifarth et al. / Atherosclerosis xxx (2012) 1e7 2. Materials and methods All procedures were approved by the institutional ethics committee and were performed according to local and federal regulations and the Declaration of Helsinki. For this study, the coronary arteries of seven donor-hearts (median age of the donors: 53, range 42e61 years) were imaged with coronary CT angiography and subsequently processed by histology. The donor-hearts were provided by the International Institute for the Advancement of Medicine (IIAM, Jessup, PA). Inclusion criteria were as follows: donor age between 40 and 70 years and history of myocardial infarction or CAD proven by diagnostic tests. The cause of death of the donorswas stroke in six cases and in one case the cause of death was non-natural (suicide). Donors, who underwent coronary artery bypass graft surgery were excluded from this study. The maximum allowed warm ischemia time was 6 h and the maximum cold ischemia time was 15 h. The fresh donor hearts were transported to our institution in Histidine-tryptophan-ketoglutarate or University of Wisconsin solution and packed in wet ice Ex vivo heart preparation and computed tomography imaging The method of donor heart preparation and ex-vivo coronary CT imaging was described previously in detail [14]. Briefly, a balloon was placed in the left ventricle and inflated with saline to maintain the physiological shape of the heart. The right and the left coronary arteries were selectively cannulated and flushed to remove air bubbles and superficial thrombi. The coronary arteries were then filled with a contrast medium (approximately 5e10 ml) consisting of low viscosity methylcellulose (Methocel, DOW Chemical Company, Midland, Michigan) and 3% of a non-ionic contrast agent (Isovue 370, Bracco Diagnostics, Milan, Italy) to achieve an intraluminal contrast enhancement similar to in-vivo coronary CT angiography (z250 HU). The organ was positioned in the center of a canola oil tank to simulate the pericardial adipose tissue. CT data acquisition was performed with a 64-detector row CT scanner (High-Definition, GE Discovery, CT 750HD) using a standard coronary CT acquisition mode and the following scan parameters: mm collimation; 0.35 s rotation time; tube voltage of 120 kv; tube current time product of 500 mas. Images were reconstructed with a slice thickness of mm. The dataset was reconstructed using an adaptive statistical iterative reconstruction technique (ASIR, GE Healthcare, Milwaukee, US). Iterative reconstruction techniques have been shown to markedly reduce the image noise and enhance plaque visualization compared to conventional image reconstruction using filtered back-projection [14]. CT images were sent to an offline workstation for further analysis. Subsequent to the CT imaging, the coronary arteries were excised along with the surrounding tissue and the side-branches were ligated. The coronaries were pressure-perfused with 10% buffered neutral formalin solution in order to achieve tissue fixation. The preparation and the CT imaging were completed within 4 h to avoid potential post-mortem changes in plaque morphology Histological preparation and classification The histological analysis was performed by experts specialized in cardiovascular pathology (CVPath Laboratory, Maryland, MD). From the coronary arteries, 6 mm thick paraffin sections were obtained in 1-mm and 5-mm increments (where minimal disease was present), resulting in a total of 611 sections. These were stained using Movat s pentachrome. Each cross-section was classified according to the modified AHA scheme [15] into the following categories: adaptive intimal thickening (AIT), pathological intimal thickening (PIT), fibrous plaque (Fib), early fibroatheromas (EFA), late fibroatheroma (LFA), thin cap fibroatheroma (TCFA). According to a report on atherosclerotic lesion classification from the American Heart Association, AIT, Fib, and PIT were considered as early atherosclerotic lesions, and EFA, LFA, and TCFA as advanced lesions [16] Co-registration of histology and computed tomography A multiplanar reconstruction technique was used to generate CT images perpendicular to the vessel centerline to match the position of the histological cuts. The cross-sectional and rotational position of the slices were identified by distance measurements from the distal end of the plastic cannulas and verified by using fiducial markers such as side branches, bifurcations and features of vessel wall morphology (e.g.: plaque shape, calcification pattern, orientations of the myocardium and the pericardial adipose tissue layer) CT data analysis As the purpose of this study was to define features of advanced atherosclerotic plaques that promote or interfere with the delineation of the NRS in coronary CT angiography, only cross-sections containing advanced plaques as classified by histology were used in this study. The corresponding CT images were reviewed in consensus by two radiologists with 6 and 10 years of experience in coronary CT angiography. All reading was performed using a fixed window setting (700 HU width, 200 HU level). Based on the CT appearance, it was specified whether a napkin-ring sign could be identified within the plaque. The NRS was defined by a lowattenuation plaque core surrounded by a circumferential area of higher attenuation [11]. Additionally, we measured the attenuation of the non-calcified plaque portion of all plaques using an in-house program developed in Matlab. For this purpose, the outer circumference of the vessel and the lumen were manually traced and the median density of the pixels within the plaque was calculated. In plaques with a positive NRS, we also measured the median density of the central hypodense area and the peripheral hyperdense ring Histologic data analysis All plaques that were previously classified as advanced were analyzed by experienced pathologists (M.N., F.O.) regarding the presence or absence of a necrotic and/or lipid core, intraplaque hemorrhage, intraplaque microvessels, macrophage infiltration and calcifications utilizing H&E and Movat s pentachrome stains. The latter characteristics were stratified as being related or unrelated to the lipid/necrotic core. Calcifications were subdivided into microcalcifications (punctate calcification of smooth muscle cells, macrophages, or extracellular matrix, maximum diameter <50 mm), spotty (coalescent calcification, maximum diameter 50e1000 mm) and sheet calcifications (plate calcification, maximum diameter > 1000 mm) [17]. All cuts were subsequently scanned and converted into high-resolution TIFF images and further analyzed using dedicated software (ImageJ 1.44o, National Institutes of Health, MD, USA). Vessel area (VA) was defined as the area inside the external elastic lamina. Furthermore, the lumen area (LA) and the area of the lipid/necrotic core (CA) were measured. From this data, the total plaque area (TPA) was calculated as TPA ¼ VA e LA. The area of noncore plaque was calculated as NcA ¼ TPA e CA. The plaque burden (PB) was calculated using the following formula: PB ¼ (TPA/ VA) * 100. We also assessed the thickness of the fibrous cap, which was measured at its thinnest part. Histologic features representing the atherosclerotic plaque such as size and area of the necrotic core and area of non-core plaque as

3 H. Seifarth et al. / Atherosclerosis xxx (2012) 1e7 3 well as the presence of calcification, hemorrhage and macrophages are further referred to as histologic plaque features. Variables associated with the plaque such as the vessel area, the lumen area, plaque burden and distance of the plaque from the ostium are further referred to as associated plaque features Statistics Continuous variables are reported as mean (standard deviation) or median [intraquartile range] and categorical by percentage (counts). In univariate analysis, we tested which parameters derived from histology were significantly different between sections with and without NRS in CT. Wilcoxon signed-rank test and Fishers exact test were used to assess for differences within continuous and categorical variables. All variables which showed a difference at a significance level of p < 0.1 were included in multivariate analysis. We fitted two different logistic regression models, one for associated variables and one for histologic variables. Non-significant variables were removed stepwise from the initial models. The remaining significant variables from both models (associated and histologic) were combined in a final multivariate logistic regression model. We derived from these models the c-statistic, which is equal to the AUC. To estimate the probability of NRS appearance in dependency of the necrotic core size, we fitted a univariate logistic regression model where the necrotic size served as the predictor. To assess the inter-observer variability of the measurements in the histology sections, the measurements were repeated by a second reader in a subset of 30 randomly chosen histology slices. The agreement between the two readers was evaluated using Pearson correlation coefficients and calculation of relative differences. A paired t-test was performed to assess whether the difference varied significantly from zero. To determine inter-observer variability for the detection of the napkin ring sign, an independent reader assessed a random subset of 100 co-registered CCTA images for the presence of the napkin ring sign. The inter-observer agreement was evaluated using Cohen s kappa statistics that was interpreted as follows: A k value greater than 0.80 corresponded to an excellent agreement and a k value of 0.61e0.80 corresponded to a good inter-observer agreement. 3. Results Histopathology identified advanced lesions (EFA, LFA and TCFA) in 139 out of the total 611 slices (22.7%). Of these, 59 (9.6%) were classified as EFA, 60 (9.8%) as LFA, and 20 (3.3%) were classified as TCFA. Early lesions (AIT, PIT and Fib) were found in 472 slices (77.3%). In CT angiography, the napkin-ring sign was observed in 33 (24%) advanced lesions, whereas no NRS was found in 106 slices (see Table 1 and Figs. 1 and 2). However, an NRS was also found in 5 of the 472 (1.1%) slices classified as early lesions. Of these five lesions, 4 were classified by histology as PIT and one as Fib. The napkin-ring sign was most commonly observed in the RCA (55%), followed by the LCX (30%) and the LAD (15%). In contrast, plaques without NRS were evenly spread among the 3 arteries. The median density of the non-calcified plaque portion was 61.8 HU [48.4e70.1 HU] for plaques with an NRS and 65.9 HU [49.3e87.7 HU] for plaques without NRS (p ¼ 0.10). In plaques with a positive NRS, the median HU of the central hypodense area was 48.1 HU [33.4e61.6 HU], whereas the rim showed a median density of 68.2 HU [52.3e76.5 HU]. The difference between the density values was significant (p < 0.001). Table 1 Distribution of histologic features in plaques with positive and negative NRS on coronary CT angiography. N NRS 33 No-NRS 106 p-value Components related with the lipid/necrotic core Macrophages 18 (55%) 48 (45%) 0.43 Angiogenesis 16 (48%) 32 (30%) 0.06 Hemorrhage 11 (33%) 25 (24%) 0.27 Micro-calcification 9 (27%) 49 (46%) 0.07 Spotty or sheet calcification 14 (42%) 31 (29%) 0.20 Components unrelated with the lipid/necrotic core Macrophages 1 (3%) 7 (7%) 0.68 Angiogenesis 4 (12%) 27 (25%) 0.15 Hemorrhage 2 (6%) 7 (7%) 1.00 Micro-calcification 2 (6%) 10 (9%) 0.73 Spotty or sheet calcification 1 (3%) 18 (17%) Histologic features corresponding to the NRS The frequency of histologic plaque features and their relation to the presence or absence of NRS on CT is summarized in Table 1. The area of the necrotic core was more than twice as large in plaques with NRS as compared to those without NRS (median 1.10 mm 2 vs mm 2, p ¼ 0.05 e see Fig. 3). Similarly, the area of non-core plaque was significantly larger in plaques with NRS as compared to those without (median mm 2 vs mm 2, p < e see Fig. 3). The thickness of the fibrous cap was not different for plaques with and without NRS (median 0.39 mm vs mm, p ¼ 0.15). Microvessels within the plaque, indicative of angiogenesis were was more common in plaques exhibiting the NRS sign (48% vs. 30%, p ¼ 0.06), while microcalcifications were more common in plaques without an NRS (27% vs. 46%, p ¼ 0.07). In contrast, the presence of macrophages, hemorrhage and calcifications (independent of whether they were spotty or sheet-like) in close proximity to the lipid core was not associated with the NRS. Additional lipid pools, smaller than the main core, were more common in plaques without an NRS (35% vs. 12%, p ¼ 0.02) Histologic plaque features distant from the core were generally not associated with the NRS except for coreunrelated spotty and sheet calcifications, which were more commonly detected in the absence of an NRS (3% vs. 17%, p ¼ 0.04). In multivariate adjusted analysis the area of the necrotic core (OR: 1.70 (95% CI: 1.12e2.57) per increase by one mm 2, p ¼ 0.01) and the area of the non-core plaque (OR: 1.24 (95% CI: 1.11e1.40) per increase by one mm 2, p < 0.01) remained as independent predictors of the NRS. In contrast, the presence of additional smaller lipid pools independently reduced the probability of the appearance of the NRS (OR: 0.22 (95% CI: 0.07e0.75) Other features associated with the NRS Plaques exhibiting an NRS in CT were located more proximally as compared to those without NRS (median distance from ostium 19.0 mm vs mm, respectively; p ¼ 0.02). Also, the vessel area was larger at the site of the plaque (median mm 2 vs mm 2, p ¼ 0.02) and plaque area itself was larger (median mm 2 vs mm 2, p < 0.001) in plaques with NRS than in plaques without NRS. Interestingly, the lumen area was not different between plaques with and without NRS (median 2.58 mm 2 vs mm 2, p ¼ 0.58). As a result, plaque burden was significantly larger in plaques with NRS as compared to those without NRS (79.8% vs. 68.5%, p < 0.001). In multivariate adjusted analysis both vessel area (OR: 0.84 (95% CI: 0.75e0.94) per 1 mm 2 increase, p < 0.01) and plaque area

4 4 H. Seifarth et al. / Atherosclerosis xxx (2012) 1e7 Fig. 1. Late fibroatheroma as classified by histology (left image) Note the large necrotic core (*) in the center of the plaque, which correlates with the hypodense center of the plaque (*) in CT (right image). The core is surrounded by prominent fibrotic tissue (open arrows), which appears as a hyperdense ring around the core in CT. Thus the plaque has a ring-like appearance in coronary CT angiography which was coined as napkin-ring sign. Additionally neovascularization is present within the plaque (closed arrow). (OR: 1.74 (95% CI: 1.36e2.21) per 1 mm 2 increase, p < ) remained independent predictors for the delineation of the NRS in CT. Plaque burden was not included in this analysis as it was co-linear and inferior to the combination of vessel area and plaque area. When adjusting for both histologic and associated features the area of the necrotic core area, the area of the non-core plaque, and the vessel area remained independent predictors for the appearance of the NRS with the area of the necrotic core being the strongest predictor (OR 1.91; 95% CI: 1.23e2.98). The combined model reached a c-statistic of (Table 2) Interobserver-agreement We found a very good correlation between the measurements for the vessel area, lumen area and the area of the necrotic core (r ¼ 0.998, and respectively) in histology. Measurements for necrotic core area and vessel area were significantly different between the two readers (p < 0.05), however, the mean difference between the measurements was 1.9% ( mm 2 ) for the vessel area and 2.9% ( mm 2 ) for the area of the necrotic core. No difference was found between the two readers for the lumen area (p ¼ 0.61). The interobserver agreement between the two CT readers to detect the napkin ring sign was excellent (Cohen s kappa ¼ 0.86; 95%CI 0.76e0.96). 4. Discussion In the search to find a pattern of coronary atherosclerotic plaque in coronary CT angiography beyond CT attenuation that is associated with advanced coronary atherosclerotic plaque several authors have described the napkin-ring sign [9,10,12]. However, the factors that influence the delineation of this sign remain unclear. In this ex-vivo study we used the gold standard histopathology to define the equivalent of the napkin-ring sign and the components of atherosclerotic plaque that correspond to this sign in coronary CT angiography. Our results demonstrate that the histopathological equivalent of the NRS is a large and advanced atherosclerotic lesion with a large necrotic core The NRS in CT and the corresponding histopathology Plaques with low density values (typically below 30HU) in coronary CT angiography correlate with lipid rich plaques as Fig. 2. Early fibroatheroma (EFA) as classified by histology (left image). Note the small lipid core in the histology image. Due to the small size, the lipid core (*) did not appear as a napkin-ring sign in coronary CT angiography.

5 H. Seifarth et al. / Atherosclerosis xxx (2012) 1e7 5 Fig. 3. Areas of the lipid/necrotic core and the non-core plaque as measured in histology in plaques with presence and absence of the napkin-ring sign in coronary CT angiography. Although there is a substantial overlap, the mean size of the core is significantly larger in plaques with positive NRS. Note that the vast majority of cores in plaques without the NRS had an area below 1 mm 2, which is deemed a crucial size often found in rupture-prone lesions. Plaques with a positive NRS in coronary CT angiography also had a significantly higher non-core plaque area as compared to plaques without the NRS. detected in IVUS or OCT [9,18]. There is also evidence, that culprit lesions in ACS show lower density values than non-culprit lesions [19]. However, due to partial volume effects, the density values show a large variability preventing a reliable differentiation of plaque types and detection of lipid-rich plaques based solely on HU values [8]. Therefore markers of atherosclerotic plaques beyond attenuation are needed for plaque stratification in CCTA. Recently, a number of studies have reported a hypodense central plaque portion surrounded by a hyperdense ring, consistent with the napkin ring sign, in patients with ACS. In the present study, the density values of the rim were significantly higher than the density measured in the center of the lesion. In a report by Pflederer et al., a similar ring like Table 2 The model combining both histologic and plaque associated features reached a c-statistic of Independent predictors OR (95%CI) p-value Vessel area [per mm 2 ] 0.88 (0.78e0.99) 0.03 Area of non-core plaque [mm 2 ] 1.60 (1.24e2.05) Area of lipid and/or necrotic core [mm 2 ] 1.91 (1.23e2.98) Presence of additional lipid cores 0.36 (0.10e1.25) 0.11 enhancement was found in 25% of culprit lesions in patients presenting with ACS but never in patients presenting with stable angina [10]. Kashiwagi et al. divided culprit lesions in patients with ACS in TCFA and non-tcfa plaques according to findings in optical coherence tomography (OCT). A ring-like enhancement was found in 44% of thin cap fibroatheromas but only in 4% of plaques without a thin fibrous cap [9]. Nishio et al. found a ring-like enhancement pattern in 44% of disrupted plaques detected by angioscopy in patients with suspected ischemic heart disease but only in 6% of plaques without signs of rupture [12]. In the present study the most distinct histopathologic feature of plaques exhibiting the NRS in CT and the strongest independent predictor was the presence of a large necrotic core representing the center of the napkin-ring sign. In plaques with positive NRS, the necrotic core was more than twice as large as in plaques without the NRS (1.10 mm 2 vs mm 2 ; respectively). Invasive studies using intravascular ultrasound (IVUS) and OCT showed that the size of the necrotic core in coronary atherosclerosis correlates with the risk for plaque rupture [5,20]. These findings have been confirmed by histopathology. Virmani et al. report in a series of 400 patients who died of sudden coronary death 80% of ruptured plaques had a necrotic core area of >1 mm 2 [2]. The second most important feature in plaques with an NRS in CT was the size of the plaque surrounding the necrotic core, which was also nearly twice as large compared to plaques without the NRS (10.15 mm 2 vs mm 2, p < 0.001). The plaque component surrounding the necrotic core mainly consists of fibrous tissue and smooth muscle cells [21] and represents the equivalent of the rim of the napkin-ring sign as seen in CT. The association between plaque size and NRS is explained in that the limited spatial resolution of CT requires a certain number of pixels representing a specific plaque component in order to be able to differentiate them from each other, i.e. the hypodense core from the hyperdense rim. While a large plaque area is required for the delineation of the napkin-ring sign, it is also a feature of advanced atherosclerotic plaques. Several CT and IVUS studies have reported that unstable lesions associated with ACS show a larger plaque area with positive remodeling compared to stable lesions in patients with stable angina [6,22,23]. The association of a large plaque area with the NRS is also mirrored in the association of NRS and plaque burden and the fact that the napkin-ring sign was more commonly detected in proximal segments of the coronary artery tree. Initially it was speculated that the ring like sign was caused by deep calcifications within the plaque. Indeed, core associated, larger calcifications such as spotty calcifications were slightly more common in plaques where the NRS was present (42% vs. 29%). However, the formation of these calcifications did not explain the appearance of the NRS. In addition, several authors demonstrated the absence of macrocalcifications around the hypodense central part of the napkin-ring sign in non-enhanced CT scans [11]. Consequently, most authors support the hypothesis that the napkin-ring sign is caused by vasa vasorum enhancing the outer curvature of the plaque. Our results demonstrate that angiogenesis is associated with the NRS (48%) but it is also quite common in plaques not characterized by an NRS (30%). In our ex-vivo study design, contrast material was viscous due to the use of methylcellulose and probably did not enter the vasa vasorum. Thus, the contrast agent did not enhance the delineation of the hyperdense ring. This is in line with previous reports demonstrating the NRS in non-contrast enhanced images [11]. However, it is conceivable that neovascularization contributes to the delineation of the napkinring sign in-vivo by increasing the attenuation of the tissue around the core after the administration of contrast media. Independent of this, the fact that neovessels are often found in plaques with an NRS is important as neovascularization arising from adventitial vasa vasorum is common in advanced coronary

6 6 H. Seifarth et al. / Atherosclerosis xxx (2012) 1e7 atherosclerotic lesions and has been associated with intraplaque hemorrhage and subsequent plaque destabilization [24,25]. Somewhat surprisingly, we found microcalcifications in the rim surrounding the core only in a quarter of NRS plaques (27%). Microcalcifications have also been suggested as a possible cause for an increase in CT attenuation as in some cases the brighter rim of the NRS has been observed in non-contrast scans [11]. Indeed, data from atherosclerosis models shows that microcalcifications often develop around the necrotic core [26]. In the present study, microcalcifications were more commonly found in plaques without an NRS (46%), suggesting that microcalcifications can elevate the Hounsfield values of the tissue around necrotic core and thus lead to volume averaging which prevents the hypodense core from being identified in coronary CT angiography. The presence of microcalcifications around the core might also explain why density values in advanced atherosclerotic plaques vary, making a reliable differentiation of plaques types impossible [8] Limitations Although we injected a contrast agent into the coronary arteries, we presume that this did not reach the small neovessels within the plaques because we used a viscous contrast agent and a low pressure to push the contrast agent into the coronary arteries. The second reason is that residual blood clots in the vasa vasorum prevent the contrast agent from entering the neovessels. In patients, the contrast agent will reach the neovessels during the first pass leading to an enhancement of the fibrous tissue around the necrotic core in advanced coronary atherosclerotic plaques. Thus, the number of plaques with a positive napkin-ring sign might be higher than in the present study. However Pflederer et al. report a similar percentage of plaques with contrast enhanced rims detected in culprit lesions in patients with ACS [1]. The current study was performed in an ex-vivo setting offering ideal conditions for the imaging of the coronary vessels, especially with regard to motion artifacts. However coronary CT angiography can be performed with little or no motion artifacts using moderns scanners careful planning of the exam [27,28]. For the present study, we used a scan protocol with a tube voltage of 120 kvp. While this has been the standard setting for many years, scan protocols using 100 kvp are now standard in smaller patients [29]. The choice of a scan protocol with lower kvp values alters the Hounsfield-units of all structures containing iodine and calcium whereas the density values of all other structures are not altered significantly [30]. Thus, the selection of an imaging protocol with 100 kvp might have slightly enhanced the detection of calcium. As the contrast material in the present study did not enter the plaque itself, the delineation of the necrotic core would not have been affected by such a protocol. While these factors might limit the generalizability of our findings our approach offered the unique possibility to compare the CT morphology of advanced atherosclerotic plaques to the gold standard of histopathology. 5. Conclusion Coronary atherosclerotic plaques with a napkin-ring sign in CT have a distinct histopathological appearance primarily characterized by a large necrotic core with a large fibrous component often associated with neovascularization. Because these features have been associated with advanced atherosclerotic and rupture-prone lesions in histology, the napkin-ring sign can possibly serve as a marker for advanced lesions in coronary CT angiography. Funding sources This study was supported by an unrestricted grant from GE Healthcare, Milwaukee, Wisconsin. Disclosures Dr. Seifarth was supported by a grant from Deutsche Forschungsgemeinschaft (DFG Se 2029/1-1). Acknowledgments None. References [1] Davies MJ. The pathophysiology of acute coronary syndromes. Heart 2000;83: 361e6. [2] Virmani R, Burke AP, Farb A, Kolodgie FD. Pathology of the vulnerable plaque. J Am Coll Cardiol 2006;47:C13e8. [3] Falk E, Shah PK, Fuster V. Coronary plaque disruption. Circulation 1995;92: 657e71. [4] Hamdan A, Assali A, Fuchs S, Battler A, Kornowski R. Imaging of vulnerable coronary artery plaques. Catheter Cardiovasc Interv 2007;70:65e74. [5] Jang I-K, Tearney GJ, MacNeill B, et al. In vivo characterization of coronary atherosclerotic plaque by use of optical coherence tomography. Circulation 2005;111:1551e5. [6] Hoffmann U, Moselewski F, Nieman K, et al. Noninvasive assessment of plaque morphology and composition in culprit and stable lesions in acute coronary syndrome and stable lesions in stable angina by multidetector computed tomography. J Am Coll Cardiol 2006;47:1655e62. [7] Marwan M, Taher MA, Meniawy El K, et al. In vivo CT detection of lipid-rich coronary artery atherosclerotic plaques using quantitative histogram analysis: a head to head comparison with IVUS. Atherosclerosis 2011;215:110e5. [8] Pohle K, Achenbach S, MacNeill B, et al. Characterization of non-calcified coronary atherosclerotic plaque by multi-detector row CT: comparison to IVUS. Atherosclerosis 2007;190:174e80. [9] Kashiwagi M, Tanaka A, Kitabata H, et al. Feasibility of Noninvasive assessment of thin-cap fibroatheroma by multidetector computed tomography. JACC Cardiovasc Imaging 2009;2:1412e9. [10] Pflederer T, Marwan M, Schepis T, et al. Characterization of culprit lesions in acute coronary syndromes using coronary dual-source CT angiography. Atherosclerosis 2010;211:437e44. [11] Maurovich-Horvat P, Hoffmann U, Vorpahl M, Nakano M, Virmani R, Alkadhi H. The napkin-ring sign: CT signature of high-risk coronary plaques? JACC Cardiovasc Imaging 2010;3:440e4. [12] Nishio M, Ueda Y, Matsuo K, et al. Detection of disrupted plaques by coronary CT: comparison with angioscopy. Heart 2011;97:1397e402. [13] Maurovich-Horvat P, Schlett CL, Alkadhi H et al. The napkin-ring sign Indicates advanced atherosclerotic lesions in coronary CT angiography. JACC Cardiovasc Imaging, in press. [14] Scheffel H, Stolzmann P, Schlett CL, et al. Coronary artery plaques: cardiac CT with model-based and adaptive-statistical iterative reconstruction technique. Eur J Radiol 2012;81(3):e363e9. [15] Virmani R, Kolodgie FD, Burke AP, Farb A, Schwartz SM. Lessons from sudden coronary death: a comprehensive morphological classification scheme for atherosclerotic lesions. Arterioscler Thromb Vasc Biol 2000;20:1262e75. [16] Stary HC, Chandler AB, Dinsmore RE, et al. A definition of advanced types of atherosclerotic lesions and a histological classification of atherosclerosis. A report from the committee on vascular lesions of the council on arteriosclerosis, American heart association. Circulation 1995;92:1355e74. [17] Burke AP, Kolodgie FD, Virmani R. Pathogenesis and significance of calcification in coronary atherosclerosis. In: Virmani R, Narula J, Leon MB, editors. The vulnerable atherosclerotic plaque: strategies for diagnosis and management. Malden: Wiley-Blackwell; p. 77e94. [18] Motoyama S, Kondo T, Anno H, et al. Atherosclerotic plaque characterization by 0.5-mm-slice multislice computed tomographic imaging. Circ J 2007;71:363e6. [19] Motoyama S, Kondo T, Sarai M, et al. Multislice computed tomographic characteristics of coronary lesions in acute coronary syndromes. J Am Coll Cardiol 2007;50:319e26. [20] Ino Y, Kubo T, Tanaka A, et al. Difference of culprit lesion morphologies between ST-segment elevation myocardial infarction and nonest-segment elevation acute coronary syndrome. JACC Cardiovasc Interv 2011;4:76e82. [21] Stary HC, Chandler AB, Glagov S, et al. A definition of initial, fatty streak, and intermediate lesions of atherosclerosis. A report from the committee on vascular lesions of the council on arteriosclerosis, American heart association. Circulation 1994;89:2462e78. [22] Burke AP, Virmani R, Galis Z, Haudenschild CC, Muller JE. Task force #2dwhat is the pathologic basis for new atherosclerosis imaging techniques? J Am Coll Cardiol 2003;41:1874e86.

7 H. Seifarth et al. / Atherosclerosis xxx (2012) 1e7 7 [23] Motoyama S, Sarai M, Harigaya H, et al. Computed tomographic angiography characteristics of atherosclerotic plaques subsequently resulting in acute coronary syndrome. J Am Coll Cardiol 2009;54:49e57. [24] Gössl M, Versari D, Hildebrandt HA, et al. Segmental heterogeneity of vasa vasorum neovascularization in human coronary atherosclerosis. JACC Cardiovasc Imaging 2010;3:32e40. [25] Virmani R. Atherosclerotic plaque progression and vulnerability to rupture: angiogenesis as a source of intraplaque hemorrhage. Arterioscler Thromb Vasc Biol 2005;25:2054e61. [26] Aikawa E, Nahrendorf M, Figueiredo JL, et al. Osteogenesis associates with inflammation in early-stage atherosclerosis evaluated by molecular imaging in Vivo. Circulation 2007;116:2841e50. [27] Dewey M, Vavere AL, Arbab-Zadeh A, et al. Patient characteristics as predictors of image quality and diagnostic accuracy of MDCT compared with conventional coronary angiography for detecting coronary artery stenoses: CORE-64 multicenter international trial. AJR 2010;194:93e102. [28] Seifarth H, Wienbeck S, Püsken M, et al. Optimal systolic and diastolic reconstruction windows for coronary CT angiography using dual-source CT. AJR 2007;189:1317e23. [29] Scheffel H, Alkadhi H, Leschka S, et al. Low-dose CT coronary angiography in the step-and-shoot mode: diagnostic performance. Heart 2008;94:1132e7. [30] Barreto M, Schoenhagen P, Nair A, et al. Potential of dual-energy computed tomography to characterize atherosclerotic plaque: ex vivo assessment of human coronary arteries in comparison to histology. JCCT 2008;2:234e42.